Doug Kerr
Well-known member
When we digitize an electrical waveform - or a camera image - we "sample" the phenomenon at regular intervals of time or space, as appropriate. The Nyquist-Shannon sampling theorem tells us that if the rate of sampling is greater than twice the highest frequency contained in the variation of the phenomenon, the collection of sample values completely "describes" the variation. The corollary is that from the collection of sample values, we can precisely reconstruct the whole pattern of variation (the waveform, image or whatever).
The process of sampling can be looked at as a case of amplitude modulation. Here, the phenomenon to be captured is the modulating signal. The carrier, rather than being a sine wave (as in familiar radio modulation), is a series of narrow pulses at the sampling rate.
By looking at sampling as amplitude modulation, we can use familiar understandings of the frequency content of a modulated signal to understand the frequency content of the set of sample values, the way in which a reconstruction filter can reconstruct the original phenomenon from the set of samples, and how aliasing occurs.
This outlook and the insights we can get through it are described in my new technical article. "Sampling as Modulation", available here:
http://dougkerr.net/Pumpkin#SamplingModulation
Just a warning. Carla, who copy edited this piece, tells me that of all the technical manuscripts of mine she had edited over the years, this was the dullest.
Best regards,
Doug
The process of sampling can be looked at as a case of amplitude modulation. Here, the phenomenon to be captured is the modulating signal. The carrier, rather than being a sine wave (as in familiar radio modulation), is a series of narrow pulses at the sampling rate.
By looking at sampling as amplitude modulation, we can use familiar understandings of the frequency content of a modulated signal to understand the frequency content of the set of sample values, the way in which a reconstruction filter can reconstruct the original phenomenon from the set of samples, and how aliasing occurs.
This outlook and the insights we can get through it are described in my new technical article. "Sampling as Modulation", available here:
http://dougkerr.net/Pumpkin#SamplingModulation
Just a warning. Carla, who copy edited this piece, tells me that of all the technical manuscripts of mine she had edited over the years, this was the dullest.
Best regards,
Doug